A motion control method and system for five-axis hybrid machining equipment

Abstract

The present invention discloses a motion control method and system for five-axis hybrid processing equipment, which relates to the field of processing equipment control technology, including: obtaining real-time status data of the five-axis hybrid processing equipment and performing data preprocessing; constructing a kinematic model based on deep learning to calculate the motion trajectory of each joint of the processing equipment; using a multi-objective optimization algorithm to plan the energy-optimal path; using an adaptive fuzzy control algorithm to generate joint drive signals according to the path planning results; based on an interference observer, real-time detection of external disturbances and adjustment of control parameters to achieve motion control. The present invention achieves the improvement of the path planning accuracy of the five-axis hybrid processing equipment, significant optimization of energy consumption, enhancement of anti-interference ability, and improvement of control response speed. It has the technical advantages of high precision, high efficiency, and high robustness, and is suitable for precision motion control scenarios in complex dynamic environments.

Description

Technical Field

[0001] The present invention relates to the technical field of machining equipment control, and in particular to a motion control method and system for five-axis hybrid machining equipment. Background Art

[0002] With the rapid development of industrial automation and intelligent manufacturing, robotics has become one of the core technologies in modern industrial production. Among them, hybrid robots, which combine the dual advantages of parallel robots and serial robots, have gradually occupied an important position in high-precision and high-flexibility tasks. Five-axis hybrid robots are widely used in precision assembly, complex processing, automated production lines and other fields due to their multiple degrees of freedom, high load capacity and precise motion characteristics. In recent years, with the increase in task complexity, hybrid robots have put forward higher requirements for motion control technology, such as improving motion planning accuracy, reducing energy consumption, and enhancing adaptability to external disturbances in multi-degree-of-freedom collaboration.

[0003] At present, the research focus of hybrid robot motion control is mainly on kinematic modeling, path planning and anti-interference control. Traditional methods usually perform path planning based on analytical kinematic models. Although they can meet certain accuracy requirements, with the complexity of the hybrid robot motion structure, these methods show obvious limitations when dealing with nonlinear, multi-degree-of-freedom collaborative motion. In addition, multi-objective optimization algorithms are gradually introduced into the field of robot path planning to simultaneously meet multiple optimization goals (such as minimizing energy consumption and maximizing path accuracy). At the same time, adaptive fuzzy control technology and disturbance observer technology are also gradually applied to motion control systems to cope with disturbances and dynamic changes in complex environments.

[0004] Although the above technologies have promoted the development of hybrid robot motion control to a certain extent, there are still many problems that need to be solved. First, the traditional analytical kinematic model is difficult to accurately describe the kinematic characteristics of the five-axis hybrid robot when faced with the complex series-parallel structure, resulting in limited motion planning accuracy. Secondly, the application of multi-objective optimization algorithms in path planning has not yet fully combined the dynamic constraints and physical characteristics of hybrid robots, which can easily lead to deviations or high energy consumption in the actual execution of the optimization results. In addition, although adaptive fuzzy control algorithms and disturbance observer technology can improve control accuracy and anti-disturbance performance, current technologies are still insufficient in dynamic adjustment and real-time response. For example, the setting of existing fuzzy rules mostly relies on experience and lacks the ability to adapt to real-time environmental changes. Summary of the invention

[0005] In view of the above-mentioned problems, the present invention is proposed.

[0006] Therefore, the technical problem solved by the present invention is that the existing five-axis hybrid machining equipment motion control technology has problems such as insufficient kinematic modeling accuracy, high path planning energy consumption and poor anti-interference performance.

[0007] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0008] In the first aspect, an embodiment of the present invention provides a motion control method for five-axis hybrid machining equipment, including obtaining real-time status data of the five-axis hybrid machining equipment and performing data preprocessing; constructing a kinematic model based on deep learning to calculate the motion trajectory of each joint of the machining equipment; using a multi-objective optimization algorithm to plan an energy-optimal path; using an adaptive fuzzy control algorithm to generate joint drive signals based on path planning results; and detecting external disturbances in real time based on an interference observer and adjusting control parameters to achieve motion control.

[0009] As a preferred scheme of the motion control method of the five-axis hybrid processing equipment described in the present invention, the real-time status data of the five-axis hybrid processing equipment includes the position information, speed information and external environmental interference data of each joint of the five-axis hybrid processing equipment; the external environmental interference data is collected through a sensor array, including vibration parameters, temperature parameters, humidity parameters and external force parameters.

[0010] As a preferred solution of the motion control method of the five-axis hybrid machining equipment described in the present invention, the data preprocessing includes: performing noise filtering on the position information and speed information of each joint of the five-axis hybrid machining equipment, and using the Kalman filter algorithm to remove high-frequency noise in the sensor signal; performing outlier detection and elimination on the external environment interference data, and filtering the valid data by setting upper and lower thresholds; normalizing the valid data, mapping different types of data to a unified numerical range to improve the consistency of the model calculation; interpolating and completing the missing data, and using the linear interpolation method to fill the missing data points generated during the acquisition process.

[0011] As a preferred solution of the motion control method of the five-axis hybrid machining equipment described in the present invention, wherein: the construction of the kinematic model based on deep learning includes:

[0012] According to the kinematic characteristics of the five-axis hybrid machining equipment, a positive kinematic function is established. and inverse kinematics functions , and introduce the kinematic model expression based on multi-mode decomposition, which is expressed as:

[0013] ,

[0014] in, is the angle vector of the machining equipment joint, To obtain the target position information of the end effector of the processing equipment, is the positive kinematic function, is the inverse kinematics function; is a normalization function that normalizes the output to a uniform range; It is the Box-Cox model decomposition function, which is used to dynamically adjust the nonlinear characteristics; It is a parameter that adjusts nonlinear characteristics and is used to optimize the performance of the model in the nonlinear region; is the sub-function of multi-mode decomposition, describing the Independent movement modes; is the predicted end-effector position; is the normalized range limit; is the number of decomposition modes;

[0015] Using the basic loss function Conduct training and combine external interference data , optimized as a dynamic weighted loss function :

[0016] ,

[0017] in, is the total number of samples of training data, The first The target location of each sample; is a positive kinematic function, according to the joint angle The calculated predicted position; , is the starting point and end point of the exercise time interval; is the measured end-effector position; is the external interference weight coefficient, which is used to dynamically adjust the loss function; is the external interference function, describing Environmental disturbances at all times;

[0018] Finally, a kinematic model based on deep learning is obtained through training;

[0019] The calculation of the motion trajectory of each joint of the machining equipment includes, according to the constructed deep learning model, taking the target position and speed of the desired end effector as input, and outputting the joint angle trajectory through model prediction, and the calculation formula is:

[0020] ,

[0021] in, The gradient of the model output position to the joint angle, is the desired velocity vector, and are the weighting coefficients of position error and velocity error respectively, To optimize the motion trajectory of each joint of the five-axis hybrid machining equipment.

[0022] As a preferred solution of the motion control method of the five-axis hybrid machining equipment described in the present invention, wherein: the use of a multi-objective optimization algorithm to plan the energy-optimal path includes comprehensively considering the multi-objective optimization problem of minimizing energy consumption and maximizing path planning accuracy, defining the objective function for:

[0023] ,

[0024] in, Five-axis hybrid machining equipment in time The power consumption, For the The expected position of the target point, For the The actual position of the target point, and are the weight coefficients of energy consumption and path accuracy, is the total number of target points;

[0025] According to the kinematic characteristics of the five-axis hybrid machining equipment, set the joint angular velocity and acceleration Physical constraints:

[0026] ,

[0027] in, and are the maximum values ​​of joint angular velocity and acceleration respectively;

[0028] A hybrid optimization method based on genetic algorithm and particle swarm algorithm is used to plan the optimal energy consumption path through global search and local optimization;

[0029] According to the optimization results, the time-displacement curve and energy consumption evaluation value of each joint are output.

[0030] As a preferred solution of the motion control method of the five-axis hybrid machining equipment described in the present invention, the generating of the joint drive signal includes defining input variables according to the path deviation and the speed deviation, and setting fuzzy rules; fuzzifying the input variables, and generating fuzzy output results through fuzzy reasoning; using a defuzzification method to convert the fuzzy output variables into accurate joint drive signals; and adaptively adjusting the fuzzy rules and parameters according to the real-time path changes to ensure control accuracy and response speed.

[0031] As a preferred solution of the motion control method of the five-axis hybrid machining equipment described in the present invention, the real-time detection of external disturbances and adjustment of control parameters include: constructing an interference observer model, estimating the interference torque on the joint motion based on the disturbance data; dynamically adjusting the control parameters according to the observation results, and correcting the joint drive signal by compensating for the interference torque; when the disturbance exceeds the set threshold range, triggering the control strategy optimization or alarm mechanism to ensure system stability and control accuracy.

[0032] In a second aspect, an embodiment of the present invention provides a motion control system for five-axis hybrid machining equipment, including:

[0033] Data acquisition module, used to obtain real-time status data of five-axis hybrid machining equipment and perform data preprocessing;

[0034] Kinematic modeling module, which is used to build a kinematic model based on deep learning and calculate the motion trajectory of each joint of the processing equipment;

[0035] Path planning module, used to plan the energy-optimal path using a multi-objective optimization algorithm;

[0036] An adaptive fuzzy control module, used to generate joint drive signals according to path planning results using an adaptive fuzzy control algorithm;

[0037] The disturbance observation and control adjustment module detects external disturbances in real time based on the disturbance observer and adjusts the control parameters to achieve motion control.

[0038] In a third aspect, an embodiment of the present invention provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the above method when executing the computer program.

[0039] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program implements the steps of the method described above when executed by a processor.

[0040] Beneficial effects of the present invention:

[0041] 1. By constructing a kinematic model based on deep learning, the motion trajectory calculation accuracy of five-axis hybrid machining equipment in complex task scenarios is effectively improved, overcoming the calculation limitations of traditional analytical models in high-dimensional nonlinear situations;

[0042] 2. The multi-objective optimization algorithm is used to comprehensively consider the minimization of energy consumption and the maximization of path accuracy, so that the planned path is not only energy-efficient and efficient, but also able to meet the requirements of high-precision tasks;

[0043] 3. The disturbance observer detects external disturbances in real time and dynamically adjusts control parameters, which significantly improves the motion stability and robustness of processing equipment in complex environments;

[0044] 4. Using adaptive fuzzy control algorithm, the joint drive signal is dynamically generated according to the path planning results and the actual operation status to ensure the control system's ability to respond quickly to environmental changes;

[0045] 5. Systematically integrate data preprocessing, deep learning modeling, path planning, multi-objective optimization and anti-interference control methods to provide a new solution for the efficient operation of five-axis hybrid machining equipment in dynamic and complex environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative work, among which:

[0047] Figure 1 This is an overall flow chart of a motion control method for five-axis hybrid machining equipment provided in the first embodiment of the present invention. DETAILED DESCRIPTION

[0048] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0049] Example 1, reference Figure 1 , is an embodiment of the present invention, and provides a motion control method for five-axis hybrid machining equipment, comprising:

[0050] S1: Obtain the real-time status data of the five-axis hybrid machining equipment and perform data preprocessing.

[0051] The real-time status data of the five-axis hybrid machining equipment includes the position information, speed information and external environmental interference data of each joint of the five-axis hybrid machining equipment; the external environmental interference data is collected through the sensor array, including vibration parameters, temperature parameters, humidity parameters and external force parameters. The layout design of the sensor array fully considers the structural characteristics of the five-axis hybrid machining equipment, high-precision encoders and accelerometers are arranged at the joints, external interference sensors are installed at key node positions, and time synchronization of multi-sensor data is achieved through synchronous clock technology.

[0052] Data preprocessing includes: noise filtering of the position and speed information of each joint of the five-axis hybrid machining equipment, using the Kalman filter algorithm combined with the median filter to remove high-frequency and low-frequency noise in the sensor signal; outlier detection and elimination of external environmental interference data, and dynamic adjustment of the upper and lower thresholds to adapt to the needs of different scenarios; normalization of valid data, mapping different types of data to a unified numerical range to improve the consistency of model calculations; interpolation and completion of missing data, using linear interpolation methods combined with multi-resolution analysis to fill in the missing data points generated during the acquisition process.

[0053] Furthermore, the data preprocessing results can send calibration signals to the sensor system through a dynamic feedback mechanism to correct sensors with large acquisition accuracy deviations. High-performance computing hardware is used to accelerate data preprocessing tasks to ensure real-time performance, while redundant sensor design is used to improve the reliability of data acquisition. The above data acquisition and preprocessing methods provide high-quality data input for subsequent motion control, which is suitable for the motion control needs of five-axis hybrid machining equipment in complex environments.

[0054] S2: Build a kinematic model based on deep learning to calculate the motion trajectory of each joint of the machining equipment.

[0055] According to the kinematic characteristics of the five-axis hybrid machining equipment, a positive kinematic function is established. and inverse kinematics functions , and further combined with multi-mode decomposition technology and deep learning algorithm to optimize the kinematic model. The model is described by the following improved expression:

[0056] ,

[0057] in, is the predicted end-effector position; is the normalization function, and its expression is:

[0058] ,

[0059] The normalization function is used to normalize the model output to a uniform numerical range. It is the normalized range limit, which is used to improve the calculation consistency of different feature quantities.

[0060] function is the Box-Cox model decomposition function, which is specifically defined as follows:

[0061] ,

[0062] in, It is a parameter that adjusts nonlinear characteristics and is used to optimize the performance of the model in the nonlinear region; As a sub-function of multi-mode decomposition, nonlinear characteristic functions (such as , or , the model's ability to fit complex kinematic characteristics is enhanced by superimposing multiple modes.

[0063] To train the model, the following basic loss function is introduced:

[0064] ,

[0065] This loss function is used to measure the deviation between the model prediction value and the true value, and combines the external disturbance data , further optimized into a dynamic weighted loss function:

[0066] ,

[0067] in, represents the measured end-effector position, is the perturbed data, is a weight parameter used to dynamically balance the impact of disturbances.

[0068] After the model training is completed, the motion trajectory of each joint of the processing equipment is calculated, including the following steps: and expected speed As input, the angular trajectory of each joint is output by optimizing and solving the following formula:

[0069] ,

[0070] in, represents the gradient of the model output position to the joint angle, and are the weight coefficients of the loss term, which are used to balance the position error and velocity error.

[0071] It should be noted that traditional kinematic modeling mostly uses analytical methods, which are computationally complex and prone to failure in high-dimensional nonlinear environments. The present invention introduces multi-mode decomposition through a deep learning model and dynamically optimizes nonlinear characteristics in combination with the Box-Cox function, greatly improving the ability to fit complex kinematic relationships.

[0072] Furthermore, external disturbance data is introduced into the loss function, and the sensitivity of the model to the disturbance data is adjusted by dynamic weights, thereby realizing the adaptive adjustment of the model in a dynamic environment. The performance of the existing model is significantly reduced in a dynamic disturbance environment, and it lacks the ability to adapt to disturbances. The present invention adds disturbance observation data items to the loss function, so that the model can weigh the disturbance in real time, thereby improving the robustness of motion control.

[0073] Furthermore, when calculating the motion trajectory of each joint of the machining equipment, the position error ( and ) and speed error ( ), by dynamically adjusting the weights and , achieving a precise balance between position and speed. Traditional methods are mostly based on single-objective optimization, which makes it difficult to balance the accuracy and dynamic characteristics of the trajectory. The present invention optimizes the weight adjustment strategy to make the trajectory calculation not only accurate but also have a high dynamic response capability.

[0074] It should also be noted that the positive kinematics function describes the mapping relationship from the joint space of the machining equipment to the workspace:

[0075] ,

[0076] in, is the position of the end effector, is the coordinate, is the direction angle, is the angle vector of each joint of the machining equipment

[0077] For five-axis hybrid machining equipment, the positive kinematics function It is usually derived from geometric constraints and degree of freedom relationships, involving the coupled motion of the series part and the parallel part, and the rotation or translation transformation matrix of each joint needs to be considered.

[0078] Furthermore, the forward kinematics model can be established using the homogeneous transformation matrix or the DH parameter method.

[0079] The inverse kinematics function describes the mapping relationship from the workspace to the joint space:

[0080] ,

[0081] Since inverse kinematics problems are usually nonlinear and multi-solution, commonly used solution methods include:

[0082] Geometric method, applicable to simple structures, solves through geometric relationships;

[0083] Analytical method, establish analytical equations to solve;

[0084] Numerical methods, such as iterative methods, Newton-Raphson methods, least squares methods, etc., are used for complex structures;

[0085] Based on machine learning methods, the model is trained through neural networks to quickly predict joint angles.

[0086] S3: Use multi-objective optimization algorithm to plan the energy-optimal path.

[0087] Considering the multi-objective optimization problem of minimizing energy consumption and maximizing path planning accuracy, the objective function is defined as:

[0088] ,

[0089] in, Five-axis hybrid machining equipment in time The power consumption, For the The expected position of the target point is the actual location, and are the weight coefficients of energy consumption and path accuracy respectively.

[0090] It should be noted that Calculate the total energy consumption by introducing the power consumption function , which explicitly expresses the processing equipment in time The energy consumption level within the range can dynamically select a more energy-saving solution when planning the path. The formula is:

[0091] ,

[0092] in, For the Joints in time The driving torque, For the Joints in time The angular velocity of Represents the total number of joints. Through this construction, the energy consumption calculation is more accurate and can reflect the actual power consumption of each joint.

[0093] Furthermore, Measure the path accuracy by And practical point The sum of squared errors between them is used as one of the optimization objectives to ensure the accuracy of the path.

[0094] According to the kinematic characteristics of the five-axis hybrid machining equipment, set the joint angular velocity and acceleration Physical constraints:

[0095] ,

[0096] in, Indicates the joint angular velocity, which is constrained not to exceed the set maximum value , to avoid mechanical system loss or loss of control caused by excessive speed; Represents the joint acceleration, constraining it not to exceed the maximum value , to ensure a smooth movement process and avoid sudden dynamic impact.

[0097] A hybrid optimization method based on genetic algorithm and particle swarm algorithm is used to plan the optimal energy consumption path through global search and local optimization. The steps are as follows:

[0098] Initial population generation: Random generation path solutions, each of which is represented by a vector containing the joint position trajectory parameters:

[0099] ,

[0100] in, Indicates The joint position vector of each path is an optimization variable, which contains the joint angle information of the processing equipment at each time point; Indicates The path in The joint position at each time node;

[0101] Fitness function calculation: Calculate the objective function value of each path solution , as its fitness value:

[0102] ,

[0103] in, is the objective function value, indicating the path solution The fitness of the path is used to evaluate the energy consumption and accuracy performance of the path; Solution for the path In time Power consumption; and are the start and end time of the path planning respectively; and are the weight coefficients of energy consumption and path accuracy in the objective function, which are used to adjust the importance of energy consumption and accuracy. is the number of target points, For the The expected position of the target point, For the The actual position of the target point;

[0104] Selection operation: select individuals of the next generation population according to the fitness value, and select the probability Inversely proportional to the fitness value:

[0105] ,

[0106] in, is the total number of path solutions in the population;

[0107] Crossover operation: Randomly select two paths from the selected population and perform a crossover operation to generate a new path solution:

[0108] ,

[0109] in, is the weight, and are two parent paths randomly selected from the population;

[0110] Mutation operation: Randomly perturb some joint positions of the new path solution:

[0111] ,

[0112] in, is the random disturbance in the mutation operation, which obeys the normal distribution , used to explore the solution space; is the standard deviation of the normal distribution, controlling the range of variation; Solve for the new path in The angular position of the joint.

[0113] The optimal path solution output by the genetic algorithm is used as the initial particle position, and the particle speed is initialized to a random value:

[0114] ,

[0115] in, For particles The velocity vector represents the moving direction and speed of the particle in the solution space;

[0116] Update the speed based on the particle's current speed, historical best position, and global best position

[0117] ,

[0118] in, It is the learning factor in particle swarm optimization, which is used to control the weight of particles approaching their own historical best position and the global best position; is a random number, For particles The best location in history, is the global best position.

[0119] Adjust the particle position according to the updated velocity:

[0120] ,

[0121] When the change in the objective function value of all particles is less than the set threshold, the algorithm stops and outputs the optimal path solution.

[0122] According to the optimization results, the time-displacement curve and energy consumption evaluation value of each joint are output, specifically:

[0123] According to the output path solution of the hybrid optimization algorithm , generate the time-displacement curve for each joint:

[0124] ,

[0125] in, For time Time joint The angle of represents the optimized joint trajectory; is an interpolation function used to generate continuous trajectories of joints.

[0126] Calculate the total energy consumption value by optimizing the path solution :

[0127] ,

[0128] Among them, the power consumption The specific expression is:

[0129] ,

[0130] Indicates joints The driving torque is calculated as:

[0131] ,

[0132] in, is the joint inertia, is the joint damping, is the gravitational torque, For joints In time The angular velocity at For joints In time Angular acceleration at .

[0133] It should be noted that existing path planning usually only focuses on a single goal (such as the shortest path length or the shortest time), and cannot take into account the comprehensive requirements of complex tasks. The present invention, through the design of the objective function, enables the path planning result to achieve the lowest energy consumption while meeting the task accuracy.

[0134] Introduction Expressing power consumption, the energy consumption evaluation of path planning is refined from macro energy consumption (total time, total displacement) to the actual motion energy consumption of joints. Existing energy consumption evaluation methods lack a detailed description of the actual energy consumption of joints, which may cause the optimization results to deviate from the actual task requirements. The present invention makes the energy consumption evaluation of path planning more accurate by introducing the actual values ​​of driving torque and angular velocity.

[0135] Furthermore, the global search of the genetic algorithm is combined with the local optimization of the particle swarm algorithm, which can quickly locate the global optimal path solution and further fine-tune the optimization. The traditional single optimization algorithm has the defects of slow convergence speed and easy to fall into the local optimum. The present invention uses a hybrid optimization method to have both global search capability and local precise optimization capability.

[0136] S4: Generate joint drive signals based on the path planning results using an adaptive fuzzy control algorithm.

[0137] Define input variables according to path deviation and speed deviation, and set fuzzy rules;

[0138] Use Path Deviation It represents the error between the current joint position of the machining equipment and the target position of the path planning, which is defined as:

[0139] ,

[0140] in, Indicates time The target location, Indicates time actual location.

[0141] Use speed deviation Represents the difference between the path planning target speed and the actual speed, defined as:

[0142] ,

[0143] in, represents the target speed, Indicates actual speed.

[0144] According to the path deviation and speed deviation Different combinations of , set fuzzy control rules, for example:

[0145] like is "positive" (position ahead) and If the speed is "positive" (faster), the drive signal should be "negative" (deceleration);

[0146] like is "small negative" (position lag) and If the speed is slow, the drive signal should be positive (accelerating).

[0147] Fuzzy rules are usually expressed by fuzzy control tables, as shown in Table 1:

[0148] Table 1 Fuzzy control table

[0149]

[0150] Fuzzy processing generates fuzzy output results through fuzzy reasoning.

[0151] The input variable and Converted into fuzzy language variables (such as "negative large", "negative medium", "negative small", etc.).

[0152] The membership function is used to describe the fuzziness of the input variables, for example:

[0153] ,

[0154] in, is the fuzzy membership function, is the input variable, which represents the input quantity of the fuzzy system; Here is the lower bound of the fuzzy interval, is the upper bound of the fuzzy interval.

[0155] A defuzzification method is used to transform the fuzzy output variables into accurate joint driving signals.

[0156] According to the fuzzy rules, combined with the input variables and The membership degree of is used to generate fuzzy output through fuzzy logic reasoning.

[0157] The output variable is a fuzzy language value (such as "positive big", "negative small", etc.).

[0158] The fuzzy output variables are converted into precise values, and the center of gravity (COG) method is used to calculate the joint drive signal:

[0159] ,

[0160] in, is the defuzzification result, which represents the joint driving signal generated by fuzzy reasoning; represents the membership function of the output variable, Represents the domain of the membership function.

[0161] Fuzzy rules and parameters are adaptively adjusted according to real-time path changes to ensure control accuracy and response speed.

[0162] Dynamically adjust fuzzy rules based on real-time path deviations and environmental changes. For example:

[0163] In highly dynamic scenes, increase the sensitivity of the rules (e.g., increase the weight of "positive large" or "negative large").

[0164] In low-dynamic scenarios, reduce the frequent adjustment of rules and enhance system stability.

[0165] Through genetic algorithm or particle swarm optimization algorithm, the weights and membership function parameters of fuzzy rules are optimized offline or online to further improve the control accuracy.

[0166] It should be noted that the adaptive fuzzy control algorithm combines the flexibility of traditional fuzzy control and the ability of real-time optimization. By dynamically adjusting fuzzy rules and control parameters, it can effectively deal with path deviation problems in complex dynamic environments and ensure the accuracy and real-time performance of joint drive signals.

[0167] Furthermore, the adaptive fuzzy control algorithm combined with the path planning results not only solves the nonlinear problem in complex path planning, but also provides a real-time and dynamic joint drive signal generation mechanism. Traditional fuzzy control rules are static and difficult to deal with path changes and noise interference in dynamic environments. The present invention ensures the real-time performance and control accuracy of the joint drive signal through an adaptive adjustment mechanism.

[0168] S5: Based on the disturbance observer, external disturbances are detected in real time and control parameters are adjusted to achieve motion control.

[0169] A disturbance observer model is constructed to estimate the disturbance torque on joint motion based on the disturbance data.

[0170] The disturbance observer is a tool used to monitor the external disturbance of processing equipment in real time. It can estimate the disturbance torque applied to the joints by the outside world in real time based on the joint state (such as position, velocity, acceleration) and dynamic model of the processing equipment. The disturbance observer can effectively capture unexpected disturbances in the operating environment, such as sudden external forces and vibrations.

[0171] The core of the disturbance observer is to compare the theoretical driving torque of the joint (the ideal value under no disturbance conditions) with the actual driving torque (the actual value after including the disturbance factor), so as to estimate the size and direction of the external disturbance. This real-time estimation provides a basis for subsequent control parameter adjustment and disturbance compensation.

[0172] The control parameters are dynamically adjusted according to the observation results, and the joint drive signal is corrected by compensating the interference torque.

[0173] After the disturbance observer estimates the external disturbance, the control system needs to dynamically adjust its parameters to compensate for the impact of the disturbance on the joint movement and ensure the accuracy and stability of the movement. The adjustment process includes the following steps:

[0174] The output of the drive signal is adjusted according to the observed disturbance torque. For example, when external disturbance is observed to cause joint operation deviation, the drive signal will automatically increase or decrease the corresponding compensation amount so that the joint movement can return to the expected trajectory.

[0175] Dynamically adjust the control system's gain parameters (such as position gain and speed gain) to enhance the system's ability to respond to disturbances. For example, when the disturbance intensity is large, increase the control system's response coefficient so that the system responds quickly and resumes smooth operation.

[0176] When the disturbance exceeds the set threshold range, the control strategy optimization or alarm mechanism is triggered to ensure system stability and control accuracy.

[0177] In order to further improve the stability of the system, when the disturbance observer detects that the external disturbance exceeds the set safety range, the control system will take the following measures:

[0178] Enable more robust control strategies. For example, in the event of excessive disturbance, the system can switch to a safe mode (such as deceleration) to protect mechanical systems and hardware components.

[0179] When the interference intensity continues to exceed the safety threshold, the system will automatically trigger an alarm to notify the operator to check the environmental conditions or equipment operating status to prevent system failure due to long-term interference.

[0180] It should be noted that the present invention solves the problem of insufficient adaptability to interference in traditional systems by detecting external interference in real time through an interference observer, combined with dynamic adjustment of control parameters and compensation mechanisms. Traditional motion control systems are usually difficult to cope with dynamic interference in complex environments, which may cause deviation of the motion trajectory, overload or instability of the mechanical system. This technology achieves efficient interference suppression capabilities through real-time detection and dynamic compensation of interference observers, allowing the system to maintain high accuracy and high stability in complex environments.

[0181] The interference compensation mechanism effectively reduces trajectory deviations caused by external interference; the ability to dynamically adjust control parameters enables the system to adapt to disturbances of varying intensities; in the event of severe interference, the protection mode and alarm mechanism are triggered, effectively extending the life of the equipment.

[0182] Embodiment 2 is the second embodiment of the present invention, which is different from the previous embodiment in that:

[0183] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention can be essentially or partly embodied in the form of a software product that contributes to the prior art or partly in the current technical solution. The current computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc., which can store program codes.

[0184] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.

[0185] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering or, if necessary, processing in another suitable manner, and then stored in a computer memory.

[0186] Embodiment 3 is an embodiment of the present invention, which provides a motion control system for five-axis hybrid machining equipment, including a data acquisition module, a kinematic modeling module, a path planning module, an adaptive fuzzy control module and an interference observation and control adjustment module.

[0187] Data acquisition module: obtains real-time status data of five-axis hybrid machining equipment and performs data preprocessing;

[0188] Kinematic modeling module: builds a kinematic model based on deep learning to calculate the motion trajectory of each joint of the processing equipment;

[0189] Path planning module: uses multi-objective optimization algorithm to plan the energy-optimal path;

[0190] Adaptive fuzzy control module: uses adaptive fuzzy control algorithm to generate joint drive signals according to path planning results;

[0191] Disturbance observation and control adjustment module: Based on the disturbance observer, it detects external disturbances in real time and adjusts control parameters to achieve motion control.

[0192] Example 4 is an embodiment of the present invention, which provides a motion control method for five-axis hybrid machining equipment. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through economic benefit calculation and simulation / comparative experiments.

[0193] The goal of this example is to verify the superiority of the five-axis hybrid machining equipment control method based on multi-objective optimization, adaptive fuzzy control and disturbance observer in terms of path accuracy, energy efficiency and anti-interference ability. The test equipment is an industrial-grade five-axis hybrid machining equipment with high-precision sensors, and its control system includes: real-time data acquisition module, kinematic model based on deep learning, fuzzy controller and disturbance observer module.

[0194] The experimental steps are:

[0195] Real-time data collection and preprocessing:

[0196] When the processing equipment performs complex multi-degree-of-freedom path tasks, sensors are used to collect joint position, speed and external disturbance data;

[0197] The data were denoised using the Kalman filter algorithm, missing data were processed using linear interpolation, and normalized to improve consistency;

[0198] Kinematic model construction and verification:

[0199] Use the motion data of five-axis hybrid machining equipment to build a deep learning kinematic model, and use a multi-mode decomposition method to optimize modeling accuracy;

[0200] Path planning and execution:

[0201] A hybrid method of genetic algorithm and particle swarm optimization algorithm is used to plan the path, balancing energy consumption and accuracy goals.

[0202] Output the time-displacement curve of each joint for execution by the control module;

[0203] Adaptive fuzzy control and dynamic adjustment:

[0204] The fuzzy controller generates joint drive signals according to path deviation and speed deviation, and adjusts the fuzzy rules in real time to improve the response speed;

[0205] Interference observation and compensation:

[0206] Real-time detection and estimation of external disturbance torque, dynamic adjustment of control parameters and triggering of alarm when the disturbance exceeds the threshold. For specific experimental data, please refer to Table 2.

[0207] Table 2 Experimental data reference table

[0208]

[0209] Test data show that joint position error and path deviation are kept below 0.76mm and 2.49mm respectively, which are significantly lower than the average error range of traditional control methods (usually 1-2mm and 3-5mm).

[0210] This result shows that the kinematic model and multi-objective optimization algorithm based on deep learning effectively improve the accuracy of path planning.

[0211] The energy consumption of each test case ranged from 332J to 398J, which is approximately 15% to 20% lower than that of traditional control methods.

[0212] The optimization algorithm successfully minimizes energy consumption while meeting the path accuracy requirements.

[0213] When the maximum interference torque reaches 0.18 N·m, the joint position error and velocity error are still maintained below 0.76 mm and 1.45 mm / s, respectively, showing good anti-interference ability.

[0214] The disturbance observer and dynamic compensation mechanism significantly improve the system's adaptability to external disturbances.

[0215] The control response time is between 7.5ms and 9.8ms, which can meet the response requirements of industrial processing equipment for high dynamic tasks, demonstrating the real-time advantage of the adaptive fuzzy control algorithm.

[0216] Through the multi-objective optimization algorithm, the path planning accuracy and energy efficiency are both improved. The interference observer adjusts the control parameters in real time, so that the system maintains stability and robustness in complex environments. The adaptive fuzzy control algorithm ensures the real-time and accuracy of the joint drive signal, and significantly reduces the execution error. This embodiment verifies the adaptability and superiority of the present invention to five-axis hybrid machining equipment, especially in terms of path accuracy, energy consumption and anti-interference performance, which is significantly better than traditional control technology, and has important industrial application value.

[0217] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A motion control method for five-axis hybrid machining equipment, characterized in that: include: Obtain real-time status data of five-axis hybrid machining equipment and perform data preprocessing; Build a kinematic model based on deep learning to calculate the motion trajectory of each joint of the processing equipment; Use multi-objective optimization algorithms to plan energy-optimal paths; Using adaptive fuzzy control algorithm, joint drive signals are generated according to path planning results; Based on the disturbance observer, the external disturbance is detected in real time and the control parameters are adjusted to realize motion control; The construction of a kinematic model based on deep learning includes: According to the kinematic characteristics of the five-axis hybrid machining equipment, a positive kinematic function is established. and inverse kinematics functions , and introduce the kinematic model expression based on multi-mode decomposition, which is expressed as: ， in, is the angle vector of the machining equipment joint, To obtain the target position information of the end effector of the processing equipment, is the positive kinematic function, is the inverse kinematics function; is a normalization function that normalizes the output to a uniform range; It is the Box-Cox model decomposition function, which is used to dynamically adjust the nonlinear characteristics; It is a parameter that adjusts nonlinear characteristics and is used to optimize the performance of the model in the nonlinear region; is the sub-function of multi-mode decomposition, describing the Independent movement modes; is the predicted end-effector position; is the normalized range limit; is the number of decomposition modes; Using the basic loss function Conduct training and combine external interference data , optimized as a dynamic weighted loss function : ， in, is the total number of samples of training data, The first The target location of each sample; is a positive kinematic function, according to the joint angle The calculated predicted position; , is the starting point and end point of the exercise time interval; is the measured end-effector position; is the external interference weight coefficient, which is used to dynamically adjust the loss function; is the external interference function, describing Environmental disturbances at all times; Finally, a kinematic model based on deep learning is obtained through training; The calculation of the motion trajectory of each joint of the machining equipment includes, according to the constructed deep learning model, taking the target position and speed of the desired end effector as input, and outputting the joint angle trajectory through model prediction, and the calculation formula is: ， in, The gradient of the model output position to the joint angle, is the desired velocity vector, and are the weighting coefficients of position error and velocity error respectively, To optimize the motion trajectory of each joint of the five-axis hybrid machining equipment.

2. The motion control method of five-axis hybrid machining equipment according to claim 1, characterized in that: The real-time status data of the five-axis hybrid machining equipment includes: The position information, speed information and external environment interference data of each joint of the five-axis hybrid machining equipment; The external environment interference data is collected through a sensor array, including vibration parameters, temperature parameters, humidity parameters and external force parameters.

3. The motion control method of five-axis hybrid machining equipment according to claim 2, characterized in that: The data preprocessing includes: The position and speed information of each joint of the five-axis hybrid machining equipment is subjected to noise filtering, and the Kalman filter algorithm is used to remove high-frequency noise in the sensor signal; Detect and remove outliers in external environment interference data, and filter valid data by setting upper and lower thresholds; Normalize the valid data and map different types of data to a unified numerical range to improve the consistency of model calculations; Missing data are interpolated and completed, and the missing data points generated during the collection process are filled using linear interpolation method.

4. The motion control method of five-axis hybrid machining equipment according to claim 3, characterized in that: The use of a multi-objective optimization algorithm to plan an energy-optimal path includes comprehensively considering the multi-objective optimization problem of minimizing energy consumption and maximizing path planning accuracy, defining an objective function for: ， in, Five-axis hybrid machining equipment in time The power consumption, For the The expected position of the target point, For the The actual position of the target point, and are the weight coefficients of energy consumption and path accuracy, is the total number of target points; According to the kinematic characteristics of the five-axis hybrid machining equipment, set the joint angular velocity and acceleration Physical constraints: ， in, and are the maximum values ​​of joint angular velocity and acceleration respectively; A hybrid optimization method based on genetic algorithm and particle swarm algorithm is used to plan the optimal energy consumption path through global search and local optimization; According to the optimization results, the time-displacement curve and energy consumption evaluation value of each joint are output.

5. The motion control method of five-axis hybrid machining equipment according to claim 4, characterized in that: Generating a joint drive signal comprises: Define input variables according to path deviation and speed deviation, and set fuzzy rules; Perform fuzzy processing on input variables and generate fuzzy output results through fuzzy reasoning; The defuzzification method is used to transform the fuzzy output variables into accurate joint driving signals; Fuzzy rules and parameters are adaptively adjusted according to real-time path changes to ensure control accuracy and response speed.

6. The motion control method of five-axis hybrid machining equipment according to claim 5, characterized in that: The real-time detection of external disturbances and adjustment of control parameters include: A disturbance observer model is constructed to estimate the disturbance torque on joint motion based on disturbance data; Dynamically adjust the control parameters according to the observation results and correct the joint drive signal by compensating the disturbance torque; When the disturbance exceeds the set threshold range, the control strategy optimization or alarm mechanism is triggered to ensure system stability and control accuracy.

7. A motion control system for five-axis hybrid machining equipment, used to implement the motion control method for five-axis hybrid machining equipment as claimed in any one of claims 1 to 6, characterized in that: include: Data acquisition module, used to obtain real-time status data of five-axis hybrid machining equipment and perform data preprocessing; Kinematic modeling module, which is used to build a kinematic model based on deep learning and calculate the motion trajectory of each joint of the processing equipment; Path planning module, used to plan the energy-optimal path using a multi-objective optimization algorithm; An adaptive fuzzy control module, used to generate joint drive signals according to path planning results using an adaptive fuzzy control algorithm; The disturbance observation and control adjustment module detects external disturbances in real time based on the disturbance observer and adjusts the control parameters to achieve motion control.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the motion control method for five-axis hybrid machining equipment according to any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the motion control method of the five-axis hybrid machining equipment according to any one of claims 1 to 6 are implemented.

Images